Unsolved

This post is more than 5 years old

1 Rookie

 • 

7 Posts

6280

October 26th, 2004 12:00

Suse Linux kernel messages "icmp v4 hw csum failure"

We have several PowerEdge 350s running SuSE 8.0 and I am really concerned about these messages, as they suddenly appeared on four different machines. The NICs are being reported as "Intel(R) 82559 Fast Ethernet LAN on Motherboard". The /var/log/messages file is filling up with messages like:

Oct 22 15:10:08 co-app-11 kernel: icmp v4 hw csum failure

Oct 22 15:10:48 co-app-11 last message repeated 2 times

The references in Goolge that I found indicate that this is a problem is the e100 driver, which we have. The solution is to set the module parameter to XsumRX=0 to turn off the checksumming. The thing that is really bothering me is why this started all of a sudden on multiple machines. In one case the machine has an uptime of 168 days and then 467 days in another case, so I don't see it as a problem caused by rebooting.

Could it be that there is some other machine on the network causing these problems? If so, how?

Any help is greatly appreaciated.

Regards,

Jim Mohr

 

1 Message

October 28th, 2004 21:00

Hallo,

have you checked your network equipment? If the four servers are all connected to the same switch it may be possible that your switch causes the checksum errors.


Greetings

Martin

1 Rookie

 • 

7 Posts

October 29th, 2004 05:00

That is most certainly the problem, but our network folks cannot find any specific. The really annoying thing is is suddenly stopped yesterday. It was like flipping a switch. At 11.08 yesterday morning the messsages stopped on all four machines. No one can think of anything at all at that time. Also, the machines are in the same logical network, but are seperated by about 150 yards.
No Events found!

Top